Update app.py
Browse files
app.py
CHANGED
|
@@ -3,24 +3,25 @@ import gradio as gr
|
|
| 3 |
|
| 4 |
from PIL import Image
|
| 5 |
import os
|
|
|
|
|
|
|
|
|
|
| 6 |
token = os.environ.get('HF_TOKEN')
|
| 7 |
whisper_to_gpt = gr.Blocks.load(name="spaces/fffiloni/whisper-to-chatGPT")
|
| 8 |
tts = gr.Interface.load(name="spaces/Flux9665/IMS-Toucan")
|
| 9 |
talking_face = gr.Blocks.load(name="spaces/fffiloni/one-shot-talking-face", api_key=token)
|
| 10 |
|
| 11 |
-
|
| 12 |
-
|
| 13 |
def infer(audio):
|
| 14 |
-
|
| 15 |
#print(gpt_response)
|
| 16 |
-
audio_response = tts(
|
| 17 |
#image = Image.open(r"wise_woman_portrait.png")
|
| 18 |
portrait_link = talking_face("wise_woman_portrait.png", audio_response, fn_index=0)
|
| 19 |
|
| 20 |
-
|
| 21 |
#portrait_response = requests.get(portrait_link, headers={'Authorization': 'Bearer ' + token})
|
| 22 |
#print(portrait_response.text)
|
| 23 |
-
|
|
|
|
| 24 |
|
| 25 |
title = """
|
| 26 |
<div style="text-align: center; max-width: 500px; margin: 0 auto;">
|
|
@@ -80,15 +81,22 @@ with gr.Blocks(css=css) as demo:
|
|
| 80 |
|
| 81 |
gr.HTML(title)
|
| 82 |
|
| 83 |
-
gpt_response = gr.Video(label="Talking Portrait response")
|
| 84 |
|
| 85 |
with gr.Column(elem_id="col-container-2"):
|
| 86 |
|
| 87 |
-
record_input = gr.Audio(source="microphone",type="filepath", label="Audio input", show_label=True,elem_id="record_btn")
|
|
|
|
| 88 |
|
| 89 |
send_btn = gr.Button("Send my request !")
|
| 90 |
-
|
| 91 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
|
| 93 |
demo.queue(max_size=32, concurrency_count=20).launch(debug=True)
|
| 94 |
|
|
|
|
| 3 |
|
| 4 |
from PIL import Image
|
| 5 |
import os
|
| 6 |
+
|
| 7 |
+
from share_btn import community_icon_html, loading_icon_html, share_js
|
| 8 |
+
|
| 9 |
token = os.environ.get('HF_TOKEN')
|
| 10 |
whisper_to_gpt = gr.Blocks.load(name="spaces/fffiloni/whisper-to-chatGPT")
|
| 11 |
tts = gr.Interface.load(name="spaces/Flux9665/IMS-Toucan")
|
| 12 |
talking_face = gr.Blocks.load(name="spaces/fffiloni/one-shot-talking-face", api_key=token)
|
| 13 |
|
|
|
|
|
|
|
| 14 |
def infer(audio):
|
| 15 |
+
whisper_to_gpt_response = whisper_to_gpt(audio, "translate", fn_index=0)
|
| 16 |
#print(gpt_response)
|
| 17 |
+
audio_response = tts(whisper_to_gpt_response[1], "English Text", "English Accent", "English Speaker's Voice", fn_index=0)
|
| 18 |
#image = Image.open(r"wise_woman_portrait.png")
|
| 19 |
portrait_link = talking_face("wise_woman_portrait.png", audio_response, fn_index=0)
|
| 20 |
|
|
|
|
| 21 |
#portrait_response = requests.get(portrait_link, headers={'Authorization': 'Bearer ' + token})
|
| 22 |
#print(portrait_response.text)
|
| 23 |
+
|
| 24 |
+
return whisper_to_gpt_response[0], portrait_link, gr.update(visible=True), gr.update(visible=True), gr.update(visible=True)
|
| 25 |
|
| 26 |
title = """
|
| 27 |
<div style="text-align: center; max-width: 500px; margin: 0 auto;">
|
|
|
|
| 81 |
|
| 82 |
gr.HTML(title)
|
| 83 |
|
| 84 |
+
gpt_response = gr.Video(label="Talking Portrait response", elem_id="video_out")
|
| 85 |
|
| 86 |
with gr.Column(elem_id="col-container-2"):
|
| 87 |
|
| 88 |
+
record_input = gr.Audio(source="microphone",type="filepath", label="Audio input", show_label=True, elem_id="record_btn")
|
| 89 |
+
whisper_tr = gr.Textbox(label="whisper english translation", elem_id="text_inp")
|
| 90 |
|
| 91 |
send_btn = gr.Button("Send my request !")
|
| 92 |
+
|
| 93 |
+
with gr.Group(elem_id="share-btn-container"):
|
| 94 |
+
community_icon = gr.HTML(community_icon_html, visible=False)
|
| 95 |
+
loading_icon = gr.HTML(loading_icon_html, visible=False)
|
| 96 |
+
share_button = gr.Button("Share to community", elem_id="share-btn", visible=False)
|
| 97 |
+
|
| 98 |
+
send_btn.click(infer, inputs=[record_input], outputs=[whisper_tr, gpt_response, share_button, community_icon, loading_icon])
|
| 99 |
+
share_button.click(None, [], [], _js=share_js)
|
| 100 |
|
| 101 |
demo.queue(max_size=32, concurrency_count=20).launch(debug=True)
|
| 102 |
|